A Review of the Taxonomy and Systematics of Aigialosaurs

Netherlands Journal of Geosciences — Geologie en Mijnbouw | 84 - 3 | 221 - 229 | 2005

A review of the taxonomy and systematics of aigialosaurs

A.R. Dutchak

Department of Geological Sciences, University of Colorado at Boulder, Boulder, Colorado 80309, USA. Email: [email protected]

Manuscript received: November 2004; accepted: January 2005

Abstract

Aigialosaurs have been recognised as a group of semi-aquatic marine reptiles for over one hundred years. While the taxonomic status of aigialosaurs has changed little in the past century, the interfamilial relationships have been modified considerably making the phylogenetic relationships between aigialosaurs, mosasaurs, dolichosaurs, coniasaurs, varanids and other squamates a topic of much debate. The monophyly of the family Aigialosauridae has been contested by recent studies and remains highly questionable. The higher-level relationships of mosasauroids within Squamata remain problematic with studies placing mosasauroids outside of Varanidae, Varanoidea and even Anguimorpha. These findings conflict with earlier views that aigialosaurs (and by association mosasaurs) were closely related to Varanus. This study concludes that further descriptions of aigialosaur taxa are needed, and several key flaws need to be addressed in the data matrices that have been used in previous studies. This should facilitate the clarification of aigialosaur systematic relationships both within Mosasauroidea and Squamata.

Keywords: aigialosaurs, mosasaurs, systematics, taxonomy

| Introduction A. novaki Kramberger, 1892, Carsosaurus marchesetti Kornhuber, 1893, Opetiosaurus bucchichi Kornhuber, 1901, Proaigialosaurus The conventional characterisation of an 'aigialosaur' is that hueni Kuhn, 1958 and Haasiasaurus gittelmani (Polcyn et al., they are semi-aquatic squamates that lived in marginal marine 1999). In addition to the specimens properly described in the habitats during the early stages of the Late Cretaceous. The literature there are two more important taxa that have played first described aigialosaurs were found in the Cenomanian- integral roles in recent discussions of aigialosaur taxonomy aged rocks along the coast of the Adriatic Sea; more recently, and phylogeny. Dallasaurus turneri Bell & Polcyn, 2005 (in Bell (1997) reported on the presence, though without any previous literature as 'the Dallas aigialosaur' although the description of the animal, of an aigialosaur from North America most recent systematic analysis places it as the sister group to (this characterisation is now revised in the present volume). clidastine mosasaurs) is described in the present volume, and Recent taxonomic and systematic questions of aigialosaur 'the Trieste aigialosaur', erroneously referred to the genus nomenclature and phylogenetic relations have focused on Opetiosaurus (Calligaris 1988) and then left unnamed by which taxa are valid, who is their closest sister group within Carroll & DeBraga (1992) and later researchers, which is in the Squamata, and whether or not there is a monophyletic process of being described (A. Paid, pers. comm.). These eight Aigialosauridae (Caldwell et al, 1995; Bell, 1997), all of which specimens represent the complete dataset upon which our harkens back to a similar debate between Kornhuber (1873) understanding of aigialosaurs is based. and Kramberger (1892). It is the goal of this study to review the literature describing There are currently six published descriptions of putative and interpreting these eight specimens from the first publi aigialosaurs: Aigialosaurus dalmaticus Kramberger, 1892, cation (Kramberger, 1892) through the most recent systematic

DownloadedNetherland from https://www.cambridge.org/cores Journal of Geoscience. IP address:s — Geologi 170.106.40.139e en Mijnbou, on 29w Sep | 202184 - at3 09:56:50 | 2005, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0016774600021004 analyses (Bell & Polcyn, 2005; Polcyn & Bell, 2005). This review Suborder Dolichosauria will identify gaps in the current knowledge of aigialosaurs, and by association, weaknesses in current systematic hypotheses. Aigialosauridae Dolichosauridae By analysing the strengths and weaknesses of previous taxo nomic and systematic interpretations it will be possible to Acteosaurus Dolichosaurus determine the best starting points for future research and the Adriosaurus directions that this research should take. Pontosaurus Aigialosaurus | Review Order Squamata The family Aigialosauridae was erected by Kramberger (1892) to contain the previously described Acteosaurus von Meyer, Suborder Lepidosauria 1860 and Adriosaurus Seeley, 1881, his newly described specimens from the island of Lesina, Italy (now Hvar, Croatia) Aigialosauridae Dolichosauridae A. dalmaticus and A. novaki, and the renamed Pontosaurus (Hydrosaurus) lesinensis originally described in 1873 by Aigialosaurus Dolichosaurus Kornhuber (for a review of P. lesinensis and dolichosaur Carsosaurus Acteosaurus Opetiosaurus systematics see Pierce & Caldwell, 2004). Kramberger grouped Pontosaurus ?Mesoleptos Adriosaurus the Aigialosauridae with the Dolichosauridae (including only Dolichosaurus longicollis Owen, 1850) in the new suborder Ophiosauria (Fig. la; this name was actually preoccupied and Serpentes was emended to Dolichosauria at a later date). Kramberger Varanidae hypothesised that the Aigialosauridae were ancestral to Aigialosauridae modern lacertilians, dolichosaurs, and pythonomorphs (snakes Mosasauridae and mosasaurs). Dolichosauridae Kramberger's classification scheme was reviewed by Helodermatidae Kornhuber (1901), who determined that the members of the Glyptosauridae family Aigialosauridae did not differ significantly from extant Anniellidae monitors and thus did not merit removal from the family Anguidae Varanidae. Kornhuber (1901) argued that the 'completely Xenosauridae different shape' of the quadrate in A. dalmaticus was not sufficient cause to erect a new family, suggesting instead that quadrate shape was extremely variable across Varanidae and Fig. 1. a. The original arrangement of aigialosaurs and dolichosaurs that the differences seen in A. dalmaticus were not excep according to Kramberger (1892); b. the modified taxonomic scheme tional. Kornhuber (1901) went on to point out that if any proposed by Nopcsa (1903) separating the long-necked dolichosaurs from specimen were to be used to illustrate a transitional form the larger aigialosaurs; c. the systematic relations of anguimorph lizards, between varanids and pythonomorphs it should be not modified from Camp (1923). Kramberger's A. dalmaticus, but instead his new specimen 0. bucchichi based on its 'special, outstanding dentition' (the and tail and reduced limbs of Acteosaurus, Adriosaurus and cone-shaped dentition of 0. bucchichi appears to have been Pontosaurus were much more similar to characteristics seen in crushed, giving the teeth a more leaf-like appearance Dolichosaurus, thus meriting their placement in the family (A. Dutchak, pers. obs.). It should be noted that in addition Dolichosauridae (Fig. lb). The remaining lizards (Aigialosaurus, to contradicting Kramberger's (1892) classification scheme, Carsosaurus, Opetiosaurus and Mesoleptos zendrini) were Kornhuber (1901) also refused to acknowledge the renaming grouped together in an emended Aigialosauridae. of Pontosaurus, repeatedly referring to the specimen as Nopcsa (1908, 1923) again reviewed the relationships of Hydrosaurus throughout his paper. fossil lizards, with the latter paper being his final word on the With two totally different classification schemes in the subject. Rejecting his earlier (Nopcsa, 1903) suggestion that literature, Nopcsa (1903) was the next to review the 'Varanus- dolichosaurs and aigialosaurs were distantly related Nopcsa like lizards of Istria'. While agreeing with Kramberger (1892) determined that they should be placed in the same family. that the Aigialosauridae were sufficiently different from After some taxonomic juggling (see Nopcsa, 1923 for details) extant varanids to merit a familial distinction, Nopcsa (1903) the family Dolichosauridae was emended to include three proposed a different distribution of genera amongst the families. subfamilies: Dolichosaurinae (Acteosaurus, Adriosaurus, It was Nopcsa (1903) who recognised that the lengthy neck Pontosaurus, Dolichosaurus and the newly named Eidolosaurus),

Aigialosaurinae (Aigialosaurus, Carsosaurus and Opetiosaurus) same with two new families being added to the superfamily and the newly erected, monogeneric Mesoleptinae (Mesoleptos). Varanoidea: the Helodermatidae and the Lanthanotidae (the Nopcsa (1923) also went to great lengths to disagree with authors obviously disagreed with the classification of earlier arguments by Fejervary (1918), who suggested that the Lanthanotus as an aigialosaur by McDowell and Bogert (1954)). cranial similarities seen in aigialosaurs and mosasaurs were Russell (1967) used Camp & Allison's (1961) taxonomic scheme a result of convergence, and to state that the subfamily in his landmark publication which focused on the Aigialosaurinae contained the ancestors of the mosasaurs. In Mosasauridae but also mentioned basal mosasauroids. While addition to supporting the aigialosaur-mosasaur relationship, discussing mosasaurian ancestors, Russell suggested that they Nopcsa (1923) also suggested an aigialosaurian-like ancestor likely passed through a body-form similar to that of for snakes, based on similarities in the caudal regions of aigialosaurs, reaffirming the close evolutionary relationship of A. dalmaticus and Pachyophis woodwardi. the two groups. Russell also took issue with the suggestion by At the same time that Nopcsa (1923) was penning his final McDowell & Bogert (1954) that Lanthanotus had its origin thoughts on the subject, Camp (1923) published his classifi within the mosasauroids and instead placed them in a polytomy cation of lizards. His taxonomic groupings (Fig. lc) were in with helodermatids, varanids and 'saniwinines' as 'tertiary' general agreement with earlier works by Dollo (1904), Williston grade lizards. (1904) and Nopcsa (1923) although the details of the taxonomy Russell (1967) covered the subject of mosasaurs so varied slightly. The families Varanidae, Dolichosauridae and thoroughly that few new studies appeared on the subject, save Aigialosauridae were grouped in the superfamily Varanoidea. various new species descriptions, for the following two decades. Camp (1923) believed that aigialosaurs were descended from Aigialosaurs were left unmentioned in the literature during 'true lizards near the Varanidae' and that they were ancestral this period. to both the mosasaurs and the dolichosaurs (Camp did not The year before Russell (1967) published his manuscript, in think snakes were dolichosaur descendants, instead placing 1966, the first English translation of Hennig's (1966) 'Grundziige Serpentes as a suborder derived from a common ancestor of einer Theorie der phylogenetischen Systematik' (Phylogenetic aigialosaurs and varanids). The classification scheme devised Systematics) appeared. While the method was set out in the by Camp (1923) was used by most researchers in the field as 1950s, the use of Hennig's parsimony analysis of phylogeny the working model until it was thoroughly revised by Estes et did not become popular until personal computers became al. (1988) using computer-based parsimony methods. available to run large analyses. The first large squamate The next to examine aigialosaurs were McDowell & Bogert phylogeny to be analysed using computers was Estes et al. (1954) in their treatise on Lanthanotus borneensis. They (1988). The paper was a first attempt to put all lizard families concluded that I. borneensis was not a highly derived varanid in a systematic context using both osteological and soft tissue but instead a relict aigialosaur. This claim was based on the characters. This study resulted in a sizeable departure from similarities in the hinge of the lower jaw (while superficially the classification of Camp (1923) on the familial level. similar, the hinges differ significantly upon closer examination), Unfortunately, fossil squamates, such as mosasaurs, dolichosaurs reduced phalangeal number (their count of four phalanges on and aigialosaurs, were not included in the analysis. the fourth digit of the aigialosaur pes has been shown to be Despite being excluded from the Estes et al. (1988) erroneous; Opetiosaurus and Aigialosaurus both show five analysis, aigialosaurs did reappear in the literature in the form phalanges in this position), and shortened limbs seen in of a survey of Adriatic lizards by Calligaris (1988). This review aigialosaurs and L. borneensis. McDowell & Bogert (1954) placed added little information to that already known from much Lanthanotus in a clade with dolichosaurs, aigialosaurs and earlier in the century with the exception of mentioning a new mosasaurs but did not hypothesise any sister group relations specimen from Komen, Slovenia. The new specimen was casually within this clade. referred to the genus Opetiosaurus but was not formally After McDowell & Bogert (1954) mentioned aigialosaurs, described. The conclusions that Calligaris (1988) drew from his Kuhn (1958) described Proaigialosaurus hueni from skull review were that the taxonomy proposed by Nopcsa (1903) was fragments found at Solnhofen in southern Germany. The sufficient (Calligaris refers to Nopcsa (1923) instead of Camp description is not very thorough and the specimen has since (1923), who emended the familiar descriptions and whose been lost. Thus Proaigialosaurus made an extremely brief, and classification scheme was more widely accepted) and that not terribly useful, appearance in aigialosaur literature. Should there was little to be done with the group until further the specimen ever be relocated, a thorough redescription and specimens were discovered. detailed drawings and photographs should serve to verify the The following descriptive and interpretative studies are all original diagnosis and allow the specimen to be placed in a cladistic studies of phylogeny, the data of which are all systematic context. explicitly available for criticism in each publication. When Camp & Allison (1961) revised the earlier classification of lizards (Camp, 1923) the general arrangement remained the

Downloaded Netherlandfrom https://www.cambridge.org/cores Journal of Geoscience. IP address:s — Geologi 170.106.40.139e en Mijnbou, on 29w Sep | 202184 - at3 09:56:50 | 2005, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0016774600021004 Carroll & DeBraga (1992) DeBraga & Carroll (1993) have received less attention than another data set produced around the same time (Bell, 1993; Carroll & DeBraga (1992) published general descriptions of see below) due in part to the characters chosen by the authors. 0. bucchichi, A. dalmaticus and Calligaris' (1988) Komen speci In many cases the characters are redundant, resulting in men, which they referred to not as Opetiosaurus but instead higher weighting of certain morphological changes, or vaguely as 'the Trieste aigialosaur'. Carroll & DeBraga (1992) assumed worded (e.g., character 62: posteromedial process of coronoid that the three specimens were closely related and shared tightly/weakly sutured to prearticular; the 'strength' or 'weak similar ways of life. This meant that they were describing a ness' of a suture is impossible to interpret without quantifying generalised 'aigialosaurian-grade' body-plan. Carroll & DeBraga the mobility of the elements involved). For example, DeBraga (1992) used the information from their descriptions to code a and Carroll's (1993) characters 1, 2, 6, 9, 17, and 21 are all fifteen-character matrix for 'Aigialosauridae' (a composite of associated with a lengthening of the snout region. Characters characters from the three specimens in the study, in effect an 24 and 25 both deal with the shape of the orbital margin of assumed monophyletic group) and nine other terminal taxa in the frontal (25 is scored as straight versus concave and 26 as an attempt to determine aigialosaur relationships within straight versus convex). These characters could easily be Anguimorpha. The resulting tree (the first published phylo- condensed into a single multi-state character. Both Globidens genetic analysis to include aigialosaurs) placed aigialosaurs in (strongly convex) and Plotosaurus (slightly concave) were a polytomy with the Lanthanotus/Varanus and Cherminotus/ scored for the shape of the frontal orbital margin when in Saniwa clades. This grouping was supported by several characters both of these genera the frontal is excluded from the orbit by (the shape of the pterygopalatine suture, the degree of the prefrontal and postorbitofrontal (Bell, 1997). Numerous contact between the supraoccipital and the parietal, the problems with character definition and weighting have led to presence or absence of a notched dentary, and the size of the DeBraga & Carroll (1993) being passed over in favour of Bell's supratemporal process of the parietal) of which only the latter (1993) study of mosasauroid interrelationships. two are visible on known aigialosaur specimens. It should be noted that, while the systematic analysis was relatively cursory | Bell (1993, 1997) by some standards, this study represents the first computer- generated systematic analysis of aigialosaur relationships. Bell (1993) produced the most complete systematic analysis of mosasauroids to date (151 characters for 37 taxa) as a part of I DeBraga & Carroll (1993) his PhD dissertation (this was pared down to 142 characters for 37 taxa by the time it was published (Bell, 1997) although DeBraga & Carroll (1993) proceeded to publish a larger-scale the modifications did not change the preferred tree topology). analysis of mosasauroid and lizard systematics. Aigialosaurs Once again the focus of the analysis was not on aigialosaurs were again coded as a single terminal taxon, negating any but instead on mosasaurs, but the analysis represented the possibility of testing their monophyly. The analysis (142 first test of the monophyly of the Aigialosauridae (although it characters and 17 taxa) was designed primarily to study the was not the first analysis to test this in press - the study was internal relationships of the family Mosasauridae but the not published for four years (Bell, 1997; Fig. 2a) - it was Aigialosauridae were found to nest as the sister group to the chronologically the first to test the hypothesis). Bell (1993, Mosasauridae and the Varanidae were determined to be the 1997) did not include Carsosaurus marchesetti or any sister group to the mosasauroids (aigialosaurs and mosasaurs). dolichosaurs, but did include 'the Dallas aigialosaur', now DeBraga & Carroll (1993) concluded mosasauroids were described as Dallasaurus turneri (Bell & Polcyn, 2005). The descended from ancestral varanids and proposed 39 character results of the analysis showed aigialosaurs to be a paraphyletic shifts that had occurred in aigialosaurs following the group with Opetiosaurus being the sister taxon to all other speciation event that separated them from the lineage of mosasauroids and Aigialosaurus grouping with the basal modern varanids. Many of these characters were visible on mosasaur Halisaurus. These results were poorly supported by only a single aigialosaur specimen and some of the characters bootstrapping tests, but this may have been a result of large for which state changes are described are not visible in any of amounts of data missing from the aigialosaurian taxa. Bell the aigialosaur specimens (premaxillary tooth count, premax- (1993, 1997) represented the first rigorous test of the mono illary bar length, ossified tympanum, anteromedial and postero phyly of the family Aigialosauridae since the erection of that medial processes of the coronoid, and strength of coronoid/ taxon. It should also be noted that Bell (1993) did not find prearticular suture). This lack of information and variation any support for a varanid/mosasauroids sister-group relation among character states within Aigialosauridae was not deemed ship, contradicting the findings of Carroll & DeBraga (1992) to be a problem as they were assumed to represent a mono and DeBraga & Carroll (1993). phyletic assemblage. Bell (1993, 1997) included all mosasauroids with the excep tion of 0. bucchichi in the family Mosasauridae on the basis of

Caldwell et al. (1995) Fig. 2. a. The interrelationships of the mosasauroids, modified from Bell's (1993) hypothesis of a paraphyletic Aigialosauridae was Bell (1997) showing a paraphyletic Aigialosauridae; b. 50% majority countered when Carsosaurus marchesetti was redescribed by rule consensus tree from Caldwell (1995) showing a monophyletic Caldwell et al. (1995). While the focus of the paper was on limb Aigialosauridae. mechanics and growth, aigialosaur systematics was also dis cussed, and the analysis (66 characters and ten taxa; Fig. 2b) set (91 characters and 15 taxa) constructed by taking Bell's represented the first published phylogenetic test of aigialosaur (1993) matrix, removing a number of taxa and then deleting monophyly, two years prior to the publication of Bell's disser any characters that were phylogenetically uninformative for tation (Bell, 1997). The results of this test were less than the remaining taxa. Caldwell (1996) did make some small spectacular, with Aigialosaurus, Opetiosaurus, Carsosaurus andadjustment s to the character codings for 0. bucchichi (which the 'Trieste aigialosaur' nesting in a polytomy with the he considered congeneric with A. dalmaticus). In addition Mosasauridae (defined in this case as mosasauroids with paddle to these changes, Caldwell (1996) referred the 'Trieste like appendages). Caldwell et al. (1995) pointed out that the aigialosaur' (Carroll & DeBraga 1992) to C. marchesetti and basal polytomy with mosasaurs was caused by a single character used the new specimen to fill in some gaps in the data set. (relatively short ribs in the posterior portion of the rib cage) The phylogeny recovered by Caldwell (1996) contradicted his whereas the aigialosaurs were united (and differentiated from earlier findings (Caldwell et al., 1995) by supporting a the mosasaurs) by eight characters including: an absence of paraphyletic Aigialosauridae; hardly a surprise considering the contact between the postorbital and postfrontal, the presence characters were taken from Bell (1993). of a premaxillary foramen, and the number of presacral vertebrae. This state of affairs indicated to Caldwell et al. (1995) that | Caldwell (1999a) aigialosaurs probably represented a monophyletic assemblage. It should also be noted that, like Bell (1993), Caldwell et al. Caldwell (1999a) used a matrix that had been further pared (1995) found no support for the varanid/mosasauroid sister down from Caldwell (1996) to examine coniasaur-mosasauroid group relationship recovered by Carroll & DeBraga (1992) and relationships. Some of the 73 characters were reworded and DeBraga & Carroll (1993). the analysis included only 11 taxa (three of which were aigialosaurs) and once again a paraphyletic Aigialosauridae Caldwell (1996) was recovered. As the matrix was still based on Bell's (1993) characters and codings, and no additional aigialosaur data With several different mosasauroid phytogenies available, were added, this result was to be expected. Caldwell (1996) reviewed the hypotheses and published a data

Downloaded fromNetherland https://www.cambridge.org/cores Journal of Geoscience. IP address:s — Geotogi170.106.40.139e en Mijnbou, on 29 Sepw 2021| 84 at- 09:56:503 | 200, subject5 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0016774600021004 Caldwell (2000) taxon to 'Taxon novum YMP' (Polcyn & Bell, 2005) if this specimen lacks diagnosable characters. In addition to the Caldwell (2000) published a further modification of the Bell presence of several superfluous taxa, the authors do not (1993) matrix that recovered a monophyletic Aigialosauridae. explain why the character scoring changes for aigialosaurian This tree topology was a result of further changes to the taxa suggested by Caldwell (1996) have not been added or matrix used in Caldwell's earlier publications (Caldwell, 1996, addressed in the text. 1999a). Eight characters from Caldwell (1999a) were removed Bell & Polcyn (2005) argue that paddle-like limbs may have (these characters dealt with the fronto-parietal suture, the evolved twice or more (judging from their tree one would posterodorsal process of the maxilla, the shape of the scapula- suggest three times) in mosasauroids. This implies that coracoid suture and the composition of the appendicular paddles cannot be used as a plesiomorphic character uniting epiphyses among other features) and a single character was mosasaurs. Instead, the plesiomorphies uniting all aigialosaurs added (the number of cervical vertebrae). The resulting matrix and mosasaurs would be features of the skull as found in had 66 characters and twelve terminal taxa (with Dolichosaurus Aigialosaurus dalmaticus and Opetiosaurus bucchichi. These longicollis added to the taxa from Caldwell, 1999a). Only two features, which are clearly plesiomorphic for the entire ingroup characters supported a monophyletic Aigialosauridae to the of Bell & Polcyn (2005), might include quadrate morphology, exclusion of all other taxa: the lack of a constricted inter- dental morphology and that of the intramandibular joint. narial process of the frontal and narrow pterygoid processes Following this reasoning, it is not aigialosaurs that are on the basisphenoid. These characters proved sufficient to subsumed within the Mosasauridae, but rather that all maintain a monophyletic Aigialosauridae even when a strict mosasauroids are subsumed within the Aigialosauridae, consensus of the 27 most parsimonious cladograms was leaving the Mosasauridae as a polyphyletic taxon. Mosasaurs constructed. It should be noted that the cranial morphology are nothing more than aigialosaurs that evolved paddle-like of C. marchesetti is unknown so the strict consensus effectively limbs at least three times in their history. Bell & Polcyn's supported the congeneric grouping of 0. bucchichi and (2005) phylogeny supports the monophyly of traditional A. dalmaticus suggested by Caldwell et al. (1995). mosasauroid subfamilies, e.g., Halisaurinae, Russellosaurinae and Mosasaurinae (although the families have been rearranged), | Bell & Polcyn (2005) and Polcyn & Bell (2005) but finds no support for a paddle-bearing common ancestor, distinct from an aigialosaur, that was by diagnosis of The most recent analyses of the mosasauroid ingroup are possession of paddles plus 'aigialosaur' cranial characters, a found in the present volume (Bell & Polcyn, 2005; Polcyn & mosasaur. This is an important perspective that influences the Bell, 2005). The systematic analysis in each publication is the interpretation of phylogenetic results that is not discussed by same, with Bell (1997) being the source of all but two of the Bell & Polcyn (2005) nor by Polcyn & Bell (2005). characters. This most recent analysis is the first to insert Bell & Polcyn's (2005) phylogeny provides two nomenclatural Haasiasaurus gittelmani, which was suggested to have possibilities for re-naming the clearly polyphyletic Mosasauridae. aigialosaurian affinities (Polcyn et al., 1999). Not surprisingly, The difference depends on the naming convention applied. For the tree topology recovered by Bell & Polcyn (2005) is similar Node-Based taxonomy: Aigialosauridae is monophyletic inclusive to that found in Bell (1997), although halisaurines were moved of all descendant taxa, and the name Mosasauridae should be from the basal position to a sister group relationship (along discarded in favour of Aigialosauridae with descendant clades with the 'Trieste aigialosaur') with russellosaurines. The major bearing new names at the respective nodes. For Stem-Based difference in the recent study is that the basal polytomy has taxonomy: a new name, Aigialosauroidea, inclusive of the most been resolved. Opetiosaurus and Aigialosaurus are found to be recent common ancestor of Opetiosaurus bucchichi (the most sequential sister taxa to the rest of the mosasauroids whereas basal mosasauroid in the phylogeny) and all of its descen Haasiasaurus is found to be the sister taxon to the clade of dants, and the Mosasauridae inclusive of the most recent (halisaurines (russellosaurines)), the 'Trieste aigialosaur' is the common ancestor of (Haasiasaurus + ((the 'Trieste aigialosaur' sister taxon to halisaurines and Dallasaurus (still referred to + Halisaurus) + Russellosaurinae)) and (Dallasaurus + as 'the Dallas aigialosaur' in the data matrix) is the sister Mosasaurinae). taxon to mosasaurines. Because the purpose of their studies was to analyse the | Global analyses including aigialosaurs relationships of mosasaur taxa in detail, the authors (Bell & Polcyn, 2005; Polcyn & Bell, 2005) may be excused for not While the series of mosasauroid ingroup analyses failed to reducing the number of taxa used in the matrix. However, if yield a definite answer as to the monophyly or paraphyly of the goal of the study is to examine relationships between taxa, the Aigialosauridae, several larger-scale squamate analyses then only diagnosable taxa should be included. One cannot were undertaken to investigate the relationships of various make an informative statement about the relationship of any taxa within Squamata (Lee, 1997; Caldwell, 1999b; Lee &

Netherlands Journal of Geosciences — Geologie en Mijnbouw | 84 - 3 | 2005 Downloaded from https://www.cambridge.org/core. IP address: 170.106.40.139, on 29 Sep 2021 at 09:56:50, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0016774600021004 ft

Caldwell, 2000; Lee & Scanlon, 2002; Pierce & Caldwell, 2004). / Iguania In no case were aigialosaurs the focus of these studies, but the y^y Gekkota results are nonetheless informative. ///*C^ Dibamidae //// ^ Amphisbaenia | Lee (1997) //// / Lacertiformes ySyS /// Cordylidae Lee (1997) used previous studies of squamate systematics <^/ ys/\y Scincidae (Pregill et al., 1986; Estes et al., 1988; Rieppel, 1988; DeBraga >v ///s^^/ Anguidae & Carroll, 1993) to construct a data matrix of 144 osteological xT^^ /// Xenosauridae ^\ //// Heloderma characters for 15 terminal taxa. The study focused on the v ^^Cs// Varanids relationships within Platynota and, unlike previous studies, x^/ Mosasauroidea included numerous fossil taxa (10 of the 15 taxa). The results ^ Serpentes of the systematic analysis indicated strong support (40 a. characters and a bootstrap score of 100) for a pythonomorph clade of mosasauroids and snakes. Lee (1997) determined that y Iguania varanids were the sister group to pythonomorphs. While this /// Anguids placement appears to contradict Carroll & DeBraga's (1992) and sSyS^ Varanoids DeBraga & Carroll's (1993) suggestion that mosasauroids were s^yS / Scincidae the sister taxon to varanids, snakes were not included in //// .XT^ Cordylidae either of the earlier studies so the pythonomorph clade could /^/^^^y/ \ Lacertiformes ^<^ ^v >> Dibamidae not be tested. \. ^X^ Gekkonoidea \. ^ Amphisbaenia | Lee (1998) \. y Serpentes xfz Mosasauroidea Lee (1998) (Fig. 3a) built upon his previous study (Lee, 1997) ^ Coniasaurus and produced a much larger data set (230 osteological characters for 22 taxa) including all squamate groups. Once again, when b. mosasauroids were included in the matrix they formed a well- supported pythonomorph clade with snakes. The mosasauroid- Fig. 3. a. The systematic relationships of squamates, modified from Lee snake relationship is supported by 43 characters (about 30 of (1998) showing mosasauroids as the sister group to Serpentes; b. squamate which deal with the braincase, tooth replacement and the relationships modified from a majority rule consensus tree (Caldwell, 1999b) mandible and intramandibular hinge), most of which are showing mosasauroids and coniasaurs as the sister group to Serpentes. diagnosable only on mosasaurian taxa. I Lee & Caldwell (2000) | Caldwell (1999b) Lee & Caldwell (2000) again recovered a well-resolved Caldwell (1999b) (Fig. 3b) also determined that the varanoid Pythonomorpha, this time using a modified version of Lee's relationships of mosasauroids proposed by Carroll & DeBraga (1998) data matrix. Lee & Caldwell (2000) used 258 characters (1992) were not supported (based on 95 osteological characters coded for 32 terminal taxa. This matrix was further modified and 21 terminal taxa), and that mosasauroids together with by Lee & Scanlon (2002), who reduced the number of characters coniasaurs formed the sister group to snakes (Fig. 6). (248) and added Mesoleptos zendrinii, the focal point of the Caldwell's (1999b) study was also based upon the previous paper. A third study (Pierce & Caldwell, 2004) reduced Lee & work of Estes et al. (1988) with the differences being the Caldwell's data set to 15 anguimorph taxa and coded them for addition of fossil taxa to the matrix and the removal of soft- 159 characters (the reasons for the reduction in characters are tissue characters. The (Mosauroidea, Coniasaurus) Serpentes) not explicitly stated but it can be assumed that the change is clade was supported by five unequivocal and four equivocal due at least in part to the reduction of terminal taxa that characters. The majority of these characters deal with the made many characters autapomorphic and uninformative). intramandibular hinge, although the presence of zygosphenes Pierce & Caldwell (2004) focused on the relationships of the and zygantra also supports the clade. The clade of dolichosaur Pontosaurus lesinensis within anguimorphs and mosasauroids, coniasaurs and snakes was found to be the thus paid little attention to the placement of the mosasauroids sister group to Scleroglossa (all other squamates except clade other than to note that it fell out as the sister group to iguanians). the dolichosaur-snake assemblage (thereby retrieving a monophyletic Pythonomorpha).

Downloaded fromNetherland https://www.cambridge.org/cores Journal of Geoscience. IP address:s — Geologi 170.106.40.139e en Mijnbou, on 29 Sepw | 2021 84 -at 309:56:50 | 200, 5subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0016774600021004 Rieppel & Zaher (2000) from redescriptions of Opetiosaurus bucchichi and Aigialosaurus dalmaticus and the description of the 'Trieste aigialosaur' While the findings of Caldwell (1999b) and Lee (1997) and the alongside the data from Dallasaurus and Haasiasaurus to get studies that stemmed from that data set (Lee, 1998; Lee & the most inclusive hypothesis of mosasauroid interrelationships Caldwell, 2000; Lee & Scanlon, 2002; Pierce & Caldwell, 2004) to date. appear to indicate a solid relationship between mosasauroids The question of where mosasauroids fit within Squamata and snakes, not everyone was convinced. The findings of Lee remains a hotly debated topic, with all of the recent analyses (1998) (and consequently all the studies that utilized tracing their roots back to Estes et al. (1988). The close variations of this matrix) were challenged by Rieppel & Zaher relationship of mosasauroids with snakes remains uncertain, (2000), who suggested that the dentition, braincase and intra- but the sister group relationship between varanoids and mandibular joint characters (about 30 of the 43 characters mosasauroids has been poorly supported in most recent suggested by Lee (1998)) that link snakes and mosasauroids studies. Lee's (1998) matrix will continue to provide a source together as pythonomorphs may be the result of convergence. of informative characters and is a good starting point for By modifying the character set and ingroup taxa from Lee future studies. The problems inherent in squamate systematics (1998), Rieppel & Zaher (2000) retrieved results that placed stem in large part from a lack of fossil data for numerous mosasauroids as the sister group to an amphisbaenian/dibamid/ groups (not the least of which are amphisbaenians and snake clade within Anguimorpha, which contradicted the dibamids). Until this problem is rectified it is unlikely that findings of Lee (1998), who found that, when all squamates squamate relationships will be distilled to a single robust are tested together, dibamids and amphisbaenians group well hypothesis. outside of Anguimorpha and instead form a monophyletic sister group to a gekkonid/pygopodid clade. By further I Acknowledgements manipulating the ingroup taxa and ordering of characters, Rieppel & Zaher (2000) were able to retrieve a tree topology The author would like to thank Dr. M. Caldwell for his guidance that grouped mosasauroids as a sister taxon to varanids and candid discussions on many topics, K. Smith and M. Polcyn (supporting the findings of Carroll & DeBraga (1992) and for their helpful reviews of the manuscript, A.S. Schulp and DeBraga & Carroll (1993)). This hypothesis requires acceptance Dr. J.W.M. Jagt and the Natuurhistorisch Museum Maastricht that amphisbaenians and dibamids are nested within for organising the First Mosasaur Meeting, the members of the Angiumorpha, a relationship that requires further investi UALVP for many helpful discussions and the many workers gation. The numerous modifications to character scoring, that have contributed to the understanding of aigialosaurs weighting, and ordering by Rieppel & Zaher (2000) serve as an and mosasaurs. Information for this paper was gathered in part excellent reminder that selective manipulation of ingroup thanks to University of Alberta GSA and FGSR travel funds. taxa and characters can allow a researcher to obtain almost any tree topology. | References

I Conclusions Bell, G.L. Jr., 1993. A phylogenetic revision of the Mosasauroidea (Squamata). Ph.D. dissertation, University of Texas, Austin, Texas: 1-293. The monophyly of the aigialosaurs is again being questioned Bell, G.L. Jr., 1997. A phylogenetic revision of North American and Adriatic (Bell & Polcyn, 2005; Polcyn & Bell, 2005), though what is clear Mosasauroidea. In: J.M. Callaway and E.L. Nicholls (eds): Ancient Marine now is that this is a matter of taxonomic definition. Cleary, Reptiles. Academic Press (San Diego): 281-332. redescriptions of the key taxa (Aigialosaurus dalmaticus, Bell, G.L. Jr. & Polcyn, M.J., 2005. Dallasaurus turneri, a new primitive Opetiosaurus bucchichi and 'the Trieste aigialosaur') are mosasauroid from the Middle Turonian of Texas and comments on the essential to further investigations into re-testing the most phytogeny of Mosasauridae (Squamata). In: Schulp, A.S. & Jagt, J.W.M. (eds): recent hypotheses. Bell's (1993) data matrix has proven to be Proceedings of the First Mosasaur Meeting. Netherlands Journal of the most popular tool for hypothesising mosasauroid phylo- Geosciences 84: 177-194. genies and the many modifications to the characters and taxa Caldwell, M.W., Carroll, R.L. S Kaiser, H., 1995. The pectoral girdle and included in the matrix have resulted in a very robust data set. forelimb of Carsosaurus marchesetti (Aigialosauridae), with a preliminary While the most streamlined versions of Bell's (1993) matrix phylogenetic analysis of mosasauroids and varanoids. Journal of Vertebrate (Caldwell, 1999a, 2000) may not be optimal for testing Paleontology 15: 516-531. relationships within Mosasauridae it is essential that the Caldwell, M.W., 1996. Ontogeny and phylogeny of the mesopodial skeleton in modifications to character scoring among the basal taxa be mosasauroids reptiles. Zoological Journal of the Linnean Society 116: 407-436. utilised or discussed by later researchers so as to continue Caldwell, M.W., 1999a. Description and phylogenetic relationships of a new improving the understanding of mosasauroid systematics. The species of Coniasaurus Owen, 1850 (Squamata). Journal of Vertebrate next step in this process is to insert the information gained Paleontology 19: 438-455.

Downloaded from https://www.cambridge.org/core. IP address: 170.106.40.139, on 29 Sep Netherland2021 at 09:56:50s Journa, subjectl too fthe Geoscience Cambridge sCore — termsGeologi of use,e e navailable Mijnbou atw | 84 - 3 | 2005 https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0016774600021004 Caldwell, M.W., 1999b. Squamate phytogeny and the retationships of snakes Von Meyer, H., 1860. Acteosaurus tommasinii aus dem schwarzen Kredeschiefern and mosasauroids. Zoological Journal of the Linnean Society 125: 115-147. von Comen am Karste. Palaeontgraphica 7: 223-231. Caldwell, M.W., 2000. On the aquatic squamate Dolichosaurus longicollis Owen, Nopcsa, F., 1903. Uber die varanusartigen Lacerten Istriens. Beitrage zur 1850 (Cenomanian, Upper Cretaceous), and the evolution of elongate necks Palaontologie und Geologie Osterreich-Ungams und des Orients 15: 31-42. in squamates. Journal of Vertebrate Paleontology 20: 720-735. Nopcsa, F., 1908. Zur Kenntnis der fossilen Eidechsen. Beitrage zur Palaontologie Calligaris, R., 1988. I rettili fossili degli 'Strati calcariei ittilitici di comeno' e und Geologie Osterreich-Ungams und des Orients 21: 33-62. dell'isola di lesina. Atti del Museo Civico di Storia naturale Trieste 41: 85-125. Nopcsa, F., 1923. Eidolosaurus und Pachyophis: zwei neue Neocom-Reptilien. Camp, C.L., 1923. Classification of the lizard families. Bulletin of the American Palaeontographica 65: 96-154. Museum of Natural History 48: 289-481. Owen, R., 1850. Description of the fossil reptiles of the Chalk Formation. In: Camp, C.L. & Allison, H.J., 1961. Bibliography of fossil vertebrates 1949-1953. Dixon, F. (ed.): The Geology and Fossils of the Tertiary and Cretaceous Memoirs of the Geological Society of America 84: 1-532. Formations of Sussex. Longman, Brown, Green and Longman (London): 378- Carroll, R.L. & DeSraga, M., 1992. Aigialosaurs: mid-Cretaceous varanoid 404. lizards. Journal of Vertebrate Paleontology 12: 66-86. Pierce, S.E. & Caldwell, M.W., 2004. Redescription and phylogenetic position DeBraga, M. & Carroll, R.L., 1993. The origin of mosasaurs as a model of of the Adriatic (Upper Cretaceous; Cenomanian) dolichosaur Pontosaurus macroevolutionary patterns and processes. Evolutionary Biology 27: 245-322. lesinensis (Kornhuber, 1873). Journal of Vertebrate Paleontology 24: 373-386. Dollo, I., 1904. L'origine des mosasauriens. Bulletin de la Societe beige de Polcyn, M.J. SBell, G.L., 2005. Russellosaurus coheni n. gen., n. sp., a 92 Geologie, de Paleontologie et d'Hydrologie 18: 217-222. million-year-old mosasaur from Texas (USA), and the definition of the Estes, R., de Quieroz, K. & Gauthier, J., 1988. Phylogenetic relationships parafamily Russellosaurina. In: Schulp, A.S. & Jagt, J.W.M. (eds): Proceedings within Squamata. In: Estes, R. & Pregill, G. (eds): Phylogenetic Relationships of the First Mosasaur Meeting. Netherlands Journal of Geosciences 84: of the Lizard Families. Stanford University Press (Stanford): 119-281. 321-333. Fejervary, G.J., 1918. Contributions to a monograph on fossil Varanidae and on Polcyn, M.J., Tchernov, E., & Jacobs, L.L., 1999. The Cretaceous biogeography Megalanidae. Annals of the National Museum of Hungary 16: 341-467. of the Eastern Mediterranean with a description of a new basal mosasauroid

Hennig, W.r 1966. Phylogenetic Systematics. University of Illinois Press from the 'Ein Yabrud, Israel. In: Tomida, Y., Rich, T. H., & Vickers-Rich, P. (Chicago): 263 pp. (eds): Proceedings of the Second Gondwanan Dinosaur Symposium, National Kornhuber, A.G., 1873. Uber einen neuen fossilen Saurier aus Lesina. Science Museum Monogaphs, No. 15: 259-290. Abhandlungen der kaiserlich-kbniglichen geologischen Reichsanstalt zu Pregill, G.K., Gauthier, J.A. & Greene, H.W., 1986. The evolution of helodermatid Wien 5: 75-90. squamates, with description of a new taxon and an overview of Varanoidea. Kornhuber, A.G., 1893. Carsosaurus Marchesettii, ein neuer fossiler Lacertilier Transactions of the San Diego Society of Natural History 21: 167-202. aus den Kreideschichten des Karstes bei Komen. Abhandlungen der Rieppel, O., 1988. A review of the origin of snakes. Evolutionary Biology 22: kaiserlich-koniglichen geologischen Reichsanstalt zu Wien 17(3): 1-15. 37-130. Kornhuber, A.G., 1901. Opetiosaurus Bucchichi, eine neue fossile Eidechse aus Rieppel, 0. & Zaher, H., 2000. The intramandibular joint in squamates, and the der unteren Kreide von Lesina in Dalmatien. Abhandlungen der kaiserlich- phylogenetic relationships of the fossil snake Pachyrachis pmblematicus koniglichen geologischen Reichsanstalt zu Wien 17(5): 1-24. Haas. Fieldiana (Geology), New Series 43: 1-69. Kramberger, K.G., 1892. Aigialosaurus, eine neue Eidechse aus den Kreide- Russell, D.A., 1967. Systematics and morphology of American mosasaurs. schiefern der Insel Lesina mit Riicksicht auf die bereits beschriebenen Peabody Museum of Natural History, Yale University, Bulletin 23: 1-241. Lacertiden von Comen und Lesina. Glasnik huvatskoga naravolosovnoga Seeley, H.G., 1881. On remains of a small lizard from Neocomian rocks of derstva (Societas historico-matulis croatica) u Zagrebu 7: 74-106. Comen, near Trieste, preserved in the Geological Museum of the University Kuhn, 0., 1958. Ein neuer Lacertilier aus dem frankischen Lithographieschiefer. of Vienna. Quarterly Journal of the Geological Society of London 37: 52-56. Neues Jahrbuch fur Geologie und Palaontologie, Monatshefte 1958: 475-510. Williston, S.W., 1904. The relationships and habits of the mosasaurs. Journal of Lee, M.S.Y., 1997. The phytogeny of varanoid lizards and the affinities of snakes. Geology 12: 43-51. Philosophical Transactions of the Royal Society, London: B 352: 53-91. Lee, M.S.Y., 1998. Convergent evolution and character correlation in burrowing reptiles: towards a resolution of squamate relationships. Biological Journal of the Linnean Society 65: 369-453. Lee, M.S.Y. & Caldwell, M.W., 2000. Adriosaurus and the affinities of mosasaurs, dolichosaurs, and snakes. Journal of Paleontology 74: 915-937. Lee, M.S.Y. S Scanlon, J.D., 2002. The Cretaceous marine squamate Mesoleptos and the origin of snakes. Bulletin of the Natural History Museum of London, Geology 68: 131-142. McDowell, S.D. S Bogert, CM., 1954. The systematic position of Lanthanotus and the affinities of the anguimorph lizards. Bulletin of the American Museum of Natural History 105: 1-142.